Introducing Pond
About Pond
Pond’s technical team features top talent, including the former ML architect of WeChat, former scientist at Amazon , and former founding principal data engineer of Yuga Labs, with over 40 publications in prestigious AI journals and conferences such as NeurIPS, ICML, PNAS, IEEE, Nature portfolios (Nature Human Behavior, Nature Communications), and Harvard Business Review.
Crypto’s AI Model Layer
Pond is on the cusp of redefining AI development in crypto, making model ownership and co-creation accessible to a new generation of innovators. By leveraging the openness, permissionless structure, and transparency of on-chain data, Pond’s model layer empowers developers to truly own, control, and monetize their AI models. Unlike traditional Web2 ecosystems, where developers often toil away under tech giants without ownership of their creations, Web3’s open infrastructure enables genuine model autonomy. Once deployed and connected to on-chain data, these models can operate independently, allowing developers to effortlessly capture and share revenue.
Apart from redefining model ownership for developers and users, Pond is lowering the barrier of entry to crypto AI by introducing a model layer that empowers developers across Web2 and Web3 to collaboratively create, own and monetize crypto-native AI models with model development guidance, guidelines and state-of-the-art model and data infrastructure. Pond’s platform and model layer are designed to meet pressing market needs while fostering innovation in a truly decentralized environment where even a general Web2 developer with no crypto knowledge can develop a crypto-native AI model.
While pushing the boundary of AI research for crypto with graph foundation models, Pond is also pioneering the applications of AI across crypto use cases, including trading, DeFi, security, MEV, recommendations and more yet to come. An example worth mentioning is the security model co-developed with an industry-leading security company GoPlus which achieves 92% for both precision and recall in predicting malicious addresses. In addition to that, the Zora NFT recommendation model achieves a 52.8% precision@5 which shows good promise considering a 6% precision from Amazon's early recommendation models. Pond is also working on dynamic DeFi fees in collaboration with OpenGradient and on MEV solutions with projects like Mamori and many other crypto-native use cases.
Dedicated to democratizing access to its model and data infrastructure, Pond is empowering the entire crypto ecosystem to advance AI in groundbreaking ways—from user behavior analysis to automated trading bots and address label classification. With Pond, the future of AI in Web3 is bright and full of transformative potential.
Updated 4 days ago